Performance and suitability assessment of controlled islanding methods for online WAMPAC application

https://doi.org/10.1016/j.ijepes.2016.05.013Get rights and content

Highlights

  • Essential differences of controlled islanding methods are assessed.

  • The key difference is the nature of the objective function used.

  • Objective function determines graph model, solving algorithm and time complexity.

  • Selection criteria of islanding methods for WAMPAC application are discussed.

Abstract

Controlled islanding can be used as an efficient action to prevent catastrophic blackouts. It splits a power system into a group of stable islands, according to an islanding solution that has been generated using an islanding method. A detailed understanding of these methods is a sound basis for the development of feasible controlled islanding methods for future networks. In this paper three methods (a) Ordered Binary Decision Diagram, (b) Weak Connection and (c) Spectral Clustering, are selected as typical islanding methods. The paper presents an analysis of their essential differences and their potential suitability for use as part of a practical implementation of controlled islanding. They are critically assessed and contrasted in terms of the objective function, graph model, solving algorithm and time complexity. The key difference between these methods is the nature of the objective function used: minimal power imbalance, minimal dynamic coupling or minimal power-flow disruption. A different objective function may require a different graph model and a different solution algorithm, which may result in the islanding methods having different time complexities and produce different islanding boundaries for the same network. Each method is applied to the IEEE 118-bus system to explore their strengths and weaknesses. Criteria for the selection of a suitable, practical method are discussed, as is the feasibility of Controlled Islanding as a WAMPAC application.

Introduction

A Wide-Area Monitoring System (WAMS) can be used to enable advanced protection and control strategies that can counteract the propagation of large disturbances and limit power system blackouts [1], [2]. An example of these advanced strategies is Controlled Islanding, which entails splitting a power system into a set of stable, electrically isolated islands and is an efficient corrective measure that can help power systems survive extreme contingencies, such as: un-damped oscillations, voltage collapse and cascading trips. The islands created are based on an islanding solution and different methods have been proposed for finding possible islanding solutions, i.e. which lines should be opened to split the power system into islands.

Many methods, e.g. Ordered Binary Decision Diagram (OBDD)-based methods, have been proposed to find the set of transmission lines that should be disconnected to create islands with good power-balance and coherent generators [3]. However, the computational efficiency of these methods may not allow them to satisfy the requirement of producing an islanding solution online, due to combinatorial explosion of the searching space. In an attempt to satisfy the requirement for an online solution, heuristic methods [4], [5], [6], [7], the Weak Connection method [8], [9] and the Spectral Clustering method [10], [11] have been proposed.

A significant potential cause of a wide area blackout is the unstable power swings that will develop as groups of generators within the system begin to lose synchronism with one another. These generator groups are referred to as coherent generator groups, as the generators within one group remain coherent with one another but cannot remain coherent with the generators in any other coherent group. As such, many islanding methods are designed to identify these coherent groups and create an islanding solution that will separate them from one another; consequently eliminating the oscillations that were a threatening to cause a wide area blackout.

All of the methods considered here include the separation of the coherent groups as a key part of the islanding solutions they create. However, despite having the same intent, each method will produce a different solution for the same network; these different solutions will also have different strengths and weaknesses. These differences, and their consequences, have neither been thoroughly addressed nor analyzed in the literature. Without a proper understanding of the differences inherent to each of these methods it will not be possible to make a reasoned choice between them when developing a controlled islanding scheme. Therefore, this paper seeks to aid in the development of such an understanding in the hope that it will serve as a sound basis for the development of feasible controlled islanding methods for future networks.

In this paper, the OBDD, Weak Connection and Spectral Clustering based islanding methods are assessed and contrasted in terms of their objective function, graph model, solving algorithm and time complexity. These methods are selected for this assessment because they represent a broad range of the methods that are currently available in the field of controlled islanding. They include methods that use a node weighted graph and methods that use edge weighted graph models that can be either directed or undirected (Section ‘Graph models of controlled islanding’). The islanding problem posed by these graph models can then be solved for one of the three commonly used objective functions, namely: minimum power imbalance, minimum dynamic coupling and minimum power-flow disruption. The optimum solution to this problem can be found using either an exhaustive search that returns every possible solution or an optimization algorithm that returns a single optimal, or near optimal, solution.

The main body of the paper is organized as follows: Section ‘Graph models and controlled islanding methods’ introduces the graph models used for controlled islanding and the three methods being assessed in this paper. Section ‘Theoretical assessment of the three methods’ presents a theoretical assessment and discussion of these methods. In Section ‘Simulation assessment of the three methods’, a case of the IEEE 118-bus system is presented to explore the strengths and weaknesses of each method. In Section ‘Criteria for selecting a suitable islanding method as part of a feasible WAMPAC application’, the criteria for selecting and islanding method for practical use and the feasibility of controlled islanding methods for use as part of a WAMPAC (Wide Area Monitoring, Protection and Control) application are discussed. Section ‘Conclusion’ concludes the paper.

Section snippets

Graph models and controlled islanding methods

This section outlines the graph models used for controlled islanding and then describes the OBDD, Weak Connection and Spectral Clustering based methods for determining controlled islanding solutions.

Theoretical assessment of the three methods

The three methods introduced in Section ‘Graph models and controlled islanding methods’ convert the controlled islanding problem into a graph-cut problem and then a combinational optimization problem, albeit using different graph models and objective functions. This section assesses these three methods in terms of the objective function, graph model, time complexity and solution algorithm.

Simulation assessment of the three methods

The differences shown in Table 1 will cause each method to perform differently in terms of solution accuracy, solution distribution and computational requirements. This section assesses the performance of the three different methods for these criteria. Time-domain simulations are also used to explore the effects of implementing the different islanding solutions. This assessment is performed using the IEEE 118-bus test system. The generator data for the classical generator models used in this

Criteria for selecting a suitable islanding method as part of a feasible WAMPAC application

In this section, the feasibility of controlled islanding as a WAMPAC application and the criteria for selecting an islanding method to be used by such an application are discussed.

Conclusion

This paper critically assesses the performance of three controlled islanding methods: OBDD, Weak Connection and Spectral Clustering. An assessment of this nature allows a deeper understanding of how the simplified islanding methods that have been proposed to solve the complex controlled islanding problem influences the islanding solution produced. The differences between the solutions found by a range of islanding methods to the same islanding problem have not been previously addressed in the

Acknowledgement

This work was supported by the NFSC under Grant 51477093, Taishan Scholars Program of Shandong Province, China, and the Yong Scholars Program of Shandong University.

References (16)

  • L. Liu et al.

    Slow coherency and angle modulated particle swarm optimization based islanding of large-scale power systems

    Adv Eng Inform

    (2009)
  • V. Terzija et al.

    Wide area monitoring, protection and control of future electric power networks

    Proc IEEE

    (2011)
  • S. Chakrabarti et al.

    Measurements get together

    IEEE Power Energ Mag

    (2009)
  • Q. Zhao et al.

    A study of system splitting strategies for island operation of power system: a two-phase method based on OBDDs

    IEEE Trans Power Syst

    (2003)
  • C. Wang et al.

    A novel real-time searching method for power system splitting boundary

    IEEE Trans Power Syst

    (2010)
  • A. Sen et al.

    A new min-cut problem with application to electric power network partitioning

    Eur Trans Electr Power

    (2009)
  • A. Peiravi et al.

    A fast algorithm for intentional islanding of power systems using the multilevel kernel k-means approach

    J Appl Sci

    (2009)
  • S.B. Yusof et al.

    Slow coherency based network partitioning including load buses

    IEEE Trans Power Syst

    (1993)
There are more references available in the full text version of this article.

Cited by (9)

  • An innovative mode-based coherency evaluation method for data-driven controlled islanding in power systems

    2023, Electric Power Systems Research
    Citation Excerpt :

    Authors in [21] have presented an optimization methodology for specifying the intentional islanding strategy in which graph theory-based network reduction technique is used to improve the computational efficacy of the model. In the second category, methods are placed that employ mixed integer programming to solve an optimization problem which its solution is an optimal islanding strategy [22–32]. The purpose of these optimization problems is to minimize power imbalance or dynamic coupling or power flow disruption [30].

  • On the use of Real-Time transformer temperature estimation for improving transformer operational tripping Schemes

    2022, International Journal of Electrical Power and Energy Systems
    Citation Excerpt :

    For oil-immersed transformers, reference [5] established a dynamic thermal model and performs experimental verification and reference [6] proposed a numerical prediction method for the hottest spot temperature of windings based on heat conduction theory. A failure in a transformer could result enormous economic loss, cause huge disturbances to all costumers connecting to it, and create the threat of voltage violation across multiple busbars [7,8]. The protection strategy of a transformer is complicated as it depends on its importance, winding connection, earthing, loading, operational modes and ambient environment, etc. [9,10]..

  • A review on intentional controlled islanding in smart power systems and generalized framework for ICI in microgrids

    2020, International Journal of Electrical Power and Energy Systems
    Citation Excerpt :

    Also, nowadays modern power systems have become larger and more complex and thus, the system operation becomes more complicated and less secure. In the past decades, major system blackouts have been a constant notice of this reality [3–5]. Cascade failures are usually commenced by an event like a fault or outage of an overloaded line.

View all citing articles on Scopus
View full text